RESULTS-BASED FINANCING IN MALAWI

OBJECTIVE

Utilization of critical maternal and neonatal services is

considered essential to meet internationally agreed-

upon maternal and child health objectives set forth in

the Millennium Development Goals (MDGs), the Sustainable

Development Goals (SDGs), and the Every Women Every Child

Action Movement’s Global Strategy.

This brief summarizes the effect of the RBF4MNH Initiative on two

of its primary aims:

1. To increase utilization of health services that were directly

targeted (including delivery and early neonatal care), and

2. To increase quality health services that were directly targeted

(including delivery and early neonatal care).

This brief also aims to provide insights into how the RBF4MNH

program affected care-seeking patterns and duration of stay

following childbirth across facilities in the participating districts. In

addition to incentives for healthcare workers, RBF4MNH included

cash incentives for pregnant women, conditional upon seeking a

facility-based delivery.

July 2016

This study was funded by the United States Agency for International Development under Translating Research into Action, Cooperative Agreement No. GHS-A-00-09-00015-00, and by the Norwegian Agency for Development Cooperation (NORAD) and the Royal Norwegian Embassy in Malawi. The project team includes prime recipient, University Research Co., LLC (URC), Harvard University School of Public Health (HSPH), and sub-recipient research organizations.

BRIEF 1: THE EFFECT OF RBF4MNH ON HEALTH SERVICE UTILIZATION AND HEALTH-SEEKING BEHAVIOR

THE RBF4MNH INITIATIVE

The Results-Based Financing for Maternal and Neonatal Health (RBF4MNH) Initiative was designed to increase uptake and improve quality of care during childbirth in Malawi. Implemented in 18 facilities in 2013 and expanded to 28 facilities in 2014 across Balaka, Dedza, Mchinji and Ntcheu districts, RBF4MNH entailed investments in infrastructure and equipment; the provision of financial incentives (based on achievement of pre-defined targets) for health providers in RBF facilities; and conditional cash transfers to pregnant women residing in catchment areas of intervention facilities for recovery of expenses directly related to accessing and staying at target facilities during and at least 48 hours after childbirth.

This series of briefs is meant to serve as a resource for decision makers as they craft results-based financing programs and policies in Malawi and similar settings. The briefs stem from a two-year impact evaluation conducted jointly by Heidelberg University in Germany and the College of Medicine in Malawi.

POLICY RECOMMENDATIONSIn light of the findings presented in this brief, the research team recommends the following when devising or modifying RBF programs

in this and similar contexts:

1. A thorough understanding of baseline utilization rates for maternal and neonatal health services is essential. This information will help to ensure that performance incentives target specific services for which utilization is low, especially in contexts like Malawi, where utilization at baseline was high for most maternal and neonatal services.

2. Consider the limited value of linking conditional cash transfers (CCT) to a service for which utilization is already very high. To produce meaningful behavioral change, cash transfers should be tied to a service or set of services for which utilization rates are sub-optimal, such as completing at least four antenatal care visits and/or seeking antenatal care within the first trimester of pregnancy.

3. Ensure that incentives promote adequate and appropriate referral across levels of care.

4. Consider scaling up the intervention to all Emergency Obstetric Care (EmOC) facilities in the participating districts to avoid potentially overburdening Comprehensive Emergency Obstetric Care (CEmOC) facilities with a large influx of patients from non-RBF4MNH Basic Emergency Obstetric Care (BEmOC) facilities.

5. In the context of generalized poverty, RBF approaches should favor an equitable approach to social health protection schemes like CCT programs, rather than targeting particular socioeconomic strata.

6. When revising the CCT component in Malawi’s RBF4MNH, and in the design of conditional or any cash-transfer programs in general:– Ensure efficiency in the registration process: find a more efficient alternative to verification by health surveillance assistants (HSAs).– Consider incentivizing facility staff to improve registration and disbursement of cash transfers. – Develop tighter controls on both the registration and the disbursement procedure to avoid fraud and selective discrimination.– Consider the placement of a CCT-focused staff member in each facility in order to promote a more seamless transfer of funds.– Invest in communication and education activities to counteract the false understanding that cash transfers could divert

attention from family planning efforts.

SUMMARY OF METHODS Findings presented in this brief are based on quantitative results

from three rounds of approximately 1,800 household surveys,

and complemented by qualitative research approaches (24

in-depth interviews with providers from 12 facilities, 36 in-

depth interviews and 29 focus group discussions with women

across the four intervention districts); the findings presented are

representative of providers’ and women’s opinions.

FINDINGS

THE RBF4MNH INITIATIVE DID NOT LEAD TO INCREASES IN HEALTH SERVICE UTILIZATION

In terms of maternal and early neonatal service utilization, the

evaluation detected no impact from the RBF4MNH initiative

on the use of delivery and early neonatal healthcare services.

Likewise, the evaluation found no effect of the initiative on

antenatal (ANC) or postnatal (PNC) care seeking.

One reason the evaluation found no difference in health service

utilization between RBF4MNH and comparison facilities could be

due to high rates of care-seeking for delivery and early neonatal

care services prior to the launch of RBF4MNH. When the initiative

100%

95%

90%

85%

80%

75%

70%

Figure 1: Percentage of women reporting having delivered their last child in a health facility

Baseline Endline

RBF4MNH facilities

Comparison facilities

Che

cklis

t Sco

re

Midline

was launched in 2013, utilization rates for delivery and early

neonatal services were upwards of 90%. While increases in service

utilization were observed between 2013 and 2015, these changes

were similar across all facilities, regardless of intervention status.

By 2015, the percentage of women who reported delivering in

a facility was approximately 95% (see Figure 1). Similarly, the

RBF4MNH initiative did not increase the percent of women seeking

PNC. Again, this could be due to the high rate of PNC service

utilization prior to the launch of RBF4MNH.

The RBF4MNH initiative did not improve the timing or frequency of

ANC care seeking, despite low baseline levels with much room for

improvement. While there were slight improvements in the timing

and frequency of ANC care-seeking, levels remained low and the

improvements were not found to be a result of the initiative. After two

years of implementation, only 20% of women sought an ANC visit

within the first trimester of pregnancy (see Figure 2) and only about

50% of women completed at least four ANC visits during pregnancy.

While these levels are consistent with national and regional trends,

ample room for improvement remains in Malawi. With slight

adjustments to the CCT component of the initiative, RBF4MNH has

an opportunity to affect the timing and frequency of ANC visits.

services increased in similar proportions across all socio-

economic strata, irrespective of intervention or control status,

so that the utilization gap between poor and least poor (which

was relatively small) remained unchanged over time. The

absence of meaningful differences at this level indicates that

RBF4MNH did not benefit the women who were less poor or

those living closest to the facilities more than others, as is

often the case when interventions are implemented without

defined targeting strategies. This could also be due to the fact

that women in these communities are all poor, so differences

across wealth quintiles are actually negligible.

70%

60%

50%

40%

30%

20%

10%

0%

Figure 2: Percentage of women reporting having attended their first antenatal care clinic during the first trimester of their most recent pregnancy

Baseline Endline

Che

cklis

t Sco

re

RBF4MNH facilities

Comparison facilities

Midline

Lilongwe

Balaka

Dedza

Mchinji

Ntcheu

The RBF4MNH initiative is currently implemented in four districts in Malawi: Balaka, Dedza, Mchinji, and Ntcheu.

THE RBF4MNH INITIATIVE DID NOT LEAD TO IMPROVEMENTS IN TERMS OF EQUITY

The RBF4MNH initiative did not have any effect on the

equitable utilization of services between women of different

income/wealth levels, nor among women living different

distances from the healthcare facility. Utilization of health

THE RBF4MNH INITIATIVE PRODUCED CHANGES IN DEMAND PATTERNS

An increasing number of women from control catchment areas

were referred to intervention facilities to receive delivery services.

Between 2013 and 2015, the percentage of women from control

areas who migrated to intervention facilities to deliver increased from

15% to 20%, after a peak of 25% in 2014 (before the intervention

was scaled up to an additional ten facilities). This increase can largely

be attributed to changes in referral rates among providers working in

control facilities. Referral rates remained stable at 15% in intervention

communities, while they increased to 21% in control communities.

Interviews with both women and healthcare workers indicated that

this shift in demand was largely induced by a change in healthcare

workers’ referral patterns. Specifically, women reported—and

healthcare workers confirmed—a greater propensity to refer women

away from control Basic Emergency Obstetric Care (BEmOC) facilities

towards Comprehensive Emergency Obstetric Care (CEmOC) facilities

(all district CEmOC facilities were RBF4MNH facilities) to ensure that

more women (hopefully those potentially more at risk) could receive

better care at delivery. By end-term, the proportion of women with

complications being referred from BEmOC to CEmOC facilities was

38% and 43% in intervention and in control communities respectively.

The RBF4MNH initiative was not intended to promote shifts

from BEmOC toward CEmOC facilities, but rather to improve

care at the BEmOC level and to reduce workload at the

CEmOC level. Therefore, the increase in referral rates observed

in control areas was an unexpected consequence. Changes over

time in referrals from control BEmOC to (intervention) CEmOC

facilities varied by district. Ntcheu saw an increase over time,

while Dedza saw a decrease over time. In Balaka, the proportion

remained relatively stable over the three years, while Mchinji

experienced a small increase. These marked differences indicate

different understandings and different implementation patterns

across districts. A desirable outcome was observed in Dedza,

where the intervention appears to have limited demand at

CEmOC level and retained more women at BEmOC level.

THE RBF4MNH INITIATIVE INCREASED THE DURATION OF POST-DELIVERY STAY AT INTERVENTION FACILITIES

While the proportion of women who stayed at a facility at least

48 hours after delivery1 increased in both groups, the evaluation

found that RBF4MNH facilities saw a significantly larger

increase than comparison facilities (see Figure 3). The indicator

Health provider with clients at Golomati Health Center, Malawi.

Photo credit: Uchembere Wangwiro Program

1 This is the national target, which was set so that women could be monitored for the onset of delivery-related complications for 48 hours post-delivery.

increased in RBF4MNH facilities nearly 30 percent more than in

comparison facilities.

Qualitative findings indicate that this change is mostly

attributable to conditional cash transfers, which women

described as giving them the means to afford the cost of

post-delivery stays (primarily the cost of food, but also to

purchase clothing for their babies). It was also reported that

district supervision teams reinforced the importance of keeping

women for at least 48 hours across all (intervention and control)

facilities, which may have supported the increase observed in

both groups of facilities. Despite these facilitators that assisted

more women to stay for monitoring, one quarter of women in

intervention facilities left earlier than 48 hours after delivery.

The evaluation also detected a tendency for women from control

areas to leave the health facility significantly earlier (i.e. within 24

hours of delivery) than women from intervention areas even when

80%

70%

60%

50%

40%

30%

20%

10%

Figure 3: Percentage of women reporting having stayed at the health facility for at least 48 hours after childbirth

Baseline Endline

Che

cklis

t Sco

re

RBF4MNH facilities

Comparison facilities

Midline

THE CONDITIONAL CASH TRANSFER WITHIN THE RBF4MNH INITIATIVE

The RBF4MNH initiative included a CCT component,

intended to motivate women to deliver in a healthcare

facility by offering financial incentives and financial

protection. The transfers included a flat lump sum and an

adjustable sum, depending on distance from the health

facility and duration of stay. Among women who received

the transfer, the average sum received was 4500 Malawian

Kwacha (approximately USD 7.15/EUR 6.53).

A health care worker in Malawi prepares to see a patientPhoto credit: SC4CCM/JSI

delivering at a RBF4MHF facility. This difference was particularly

pronounced at end-term. It is likely that this difference is motivated

by the potential of the CCT (conditional on the 48-hrs post-

partum stay) to retain intervention women in facilities. This finding

provides evidence that the removal of geographical targeting for

the CCT could result in longer length of stay for all women, thereby

expanding the health benefits of the program.

THE INTERVENTION DID NOT AFFECT PATTERNS OF OUT-OF-POCKET SPENDING FOR DELIVERY SERVICES

The evaluation did not detect any difference in the amount

paid by women to cover direct costs (transport, minor medical

expenses, food, supplies for mother and baby) associated with

NON-REGISTRATION IF LATE TO INITIATE ANTENATAL CARE

“They gave cards to women who started

antenatal clinic at four months, but the

cards were not given for those women who

started antenatal at six or seven months.”

— Focus Group Discussion

delivery at a health facility. The average out-of-pocket spending

increased homogeneously across intervention and control

facilities, from 2300 MKW in 2013 to 3500 MKW in 2015. The

increase is likely to be due to inflation and to the decreasing

purchasing power of the local currency. The fact that the

evaluation did not detect an effect of the intervention on total

out-of-pocket spending indicates that healthcare providers did

not take advantage of intervention women (for example, through

the imposition of informal charges) in light of the fact that they

knew these women would be receiving a CCT. Women who

received the CCT received on average 4500 MKW, suggesting

that on average, they were amply able to recover the direct

costs incurred due to the facility-based delivery. This suggests

there is potential for the CCT to act as an instrument of financial

protection. In addition, women receiving the CCT were found to

rely less heavily on a family member accompanying them and

remaining with them at the facility for the entire duration of the

pre- and post-partum phase. This implies that the costs of care

may be lower among women receiving the CCT due to lower

productivity losses among caregivers

THE CCT PROGRAM DID NOT REACH ALL ELIGIBLE WOMEN AS INTENDED

Out of all eligible women (i.e., women residing in the catchment

areas of the RBF4MNH facilities), just above 50% were registered

in the cash transfer program. Of those, only two-thirds reported

having received the cash transfer upon delivering in the

healthcare facility. Registration rates appeared to be highest

in Balaka (70% of all eligible women registered) and lowest

in Dedza (40% of all eligible women registered). Dedza also

displayed the lowest percentage of women receiving the transfer

once registered (55%), while Ntcheu displayed the highest

percentage of women receiving the transfer once registered

(84%). Beyond the differences across districts, registration did

follow an equity-oriented pattern, with rural and least wealthy

women being the most likely to be registered in the program.

Providers and women said a major barrier to registration was

that patients were not allowed to register for cash transfers if they

visited facilities too late during the antenatal period; women who

made their first visit during their third trimester were generally not

registered (See Box: Non-Registration if Late to Initiate Antenatal

Care). In some cases, however, women who made their first

visit during the fifth month of pregnancy were also considered

ineligible for the CCT. The Initiative has since changed registration

procedures so that all eligible women are registered regardless

of the timing of the first antenatal care visit. Some women

expressed suspicion that providers discriminated against certain

women when deciding who would receive transfers, or showed

nepotism towards clients whom they knew. Providers and women

also reported challenges related to the CCT, including:

n An absence of available funds at the time of discharge after delivery

n A woman’s inability to produce evidence of her earlier registration

(i.e., a registration card missing both at the facility and among

the paperwork that women keep in their possession)

n An inability among women to prove the distance travelled

between their villages and the facility and uneven

documentation by HSAs to verify distance travelled

n Overall, poor quality and late submission by HSAs of

verification cards

n Cumbersome process of multiple disbursements weighed on

overworked providers

n Ineffective timing of disbursements to clients, hindering ability

to purchase necessary items

CASH TRANSFERS DID NOT MEASURABLY ALTER CARE-SEEKING BEHAVIORS AMONG WOMEN, BUT WOMEN REPORTED THAT THE CASH DID ENABLE THEIR CARE SEEKING

Given the joint implementation of demand-side and supply-side

incentives, the evaluation could not detect the specific effect of

CCT on utilization of maternal care services. Among women with

self-reported complications, the evaluation detected a significant

reduction in time to care (the time elapsed between onset of

symptoms and the decision to seek professional care) attributable

to the CCT. Through qualitative interviews, women generally

described the cash in a positive light, referring to it as “helpful”

(see Box: Cash Transfers as Helpful). Women reported the cash

payments positively influenced their decision to:

A health provider cares for mother and newborn at a

health center in Dedza District.

Photo credit: Uchembere Wangwiro Program

n Stay longer in facilities, given that they had money for food

and baby supplies

n Arrive earlier during labor, for the same reason

Providers observed that among women who received the cash

transfer (and the sensitization that accompanied it), women’s

empowerment and awareness of patient rights during maternity

care was stronger, especially among the poor.

WOMEN USED THE CASH RECEIVED TO PURCHASE FOOD AND BABY SUPPLIES

Women described using the funds to purchase food and items

related to caring for the baby, such as clothing, baby soap, and

an umbrella. In two instances (reported in one facility each in

Ntcheu and Mchinji districts), women described being pressured

to purchase items directly from providers. Another point of

concern related to reports from women in one facility, who felt

pressured to pay for services that were once provided free of

charge, such as baby weighing, as providers remarked that they

knew women had enough funds to pay a small fee, given the

recent receipt of a cash transfer.

CONCERNS THAT CASH TRANSFERS WOULD INCENTIVIZE WOMEN TO BECOME PREGNANT WERE NOT WARRANTED

At the outset of the program, all stakeholders expressed concern

that providing cash transfers to women may incentivize them

to become pregnant. In follow-up interviews one year into

the program, stakeholders highlighted that these fears were

unwarranted, as women were not viewing pregnancy as a

CASH TRANSFERS AS HELPFUL

“I feel the money is very helpful because

sometimes you may find that you have

been referred to the district hospital and

your husband doesn’t have any money,

the situation becomes bearable because

of the money.”

— Focus Group Discussion

University Research Co., LLC � 7200 Wisconsin Avenue, Suite 600 � Bethesda, MD 20814 � 301-654-8338 � www.urc-chs.com

TRACTION PROJECT OVERVIEWThe Translating Research Into Action (TRAction) Project, funded by the U.S. Agency for International Development, focuses on implementation and delivery science—which seeks to develop, test, and compare approaches to more effectively deliver health interventions, increase utilization, achieve coverage, and scale-up evidence-based interventions. TRAction supports implementation research to provide critically-needed evidence to program implementers and policy-makers addressing maternal and child health issues.

For more information on the TRAction Project: www.tractionproject.org � tracinfo@urc-chs.com

RBF4MNH IMPLEMENTING PARTNERSRBF4MNH is implemented by the Malawian Ministry of Health with funding from the Governments of Germany and Norway through KfW. Technical assistance to the Ministry of Health is provided through Options Consultancy Ltd. Additional information on the program can be found at:

www.options.co.uk

mechanism to access cash. Stakeholders said there were two

reasons their fears did not come to fruition. First, the amount

of funds available for cash transfers was too low to compel

pregnancy. Second, as the broader RBF4MNH program was

rolled out, it was adapted to include family planning components.

No woman described cash transfers as a factor that compelled

her to become pregnant.

Study Contact: Dr. Manuela De Allegri, Principal Investigator of the RBF4MNH Evaluationmanuela.de.allegri@urz.uni-heidelberg.de